First leaked 3DMark benchmarks show Radeon RX 7900 cards close to GeForce RTX 4080

I'm keeping an open mind about this, alleged leaked RX 7900 XTX gaming benchmarks in some games looks to sit between the RTX 4080 and RTX 4090.

For an AMD favored game like AC Valhalla, the RX 7900 XTX getting within arm's length of the RTX 4090. Guess we'll find out on the 13th, can hardly wait.

 

never seen any other leaks, these are first

this might be frequency related indeed, see how timespy dx12 vs firestrike dx11 show a big difference.





now imagine adding additional ray tracing compute to that in dxr,will go from bad to terrible.

see how that's in 4K dx12 under higher load, less in 1440p dx12 and then even less in dx11
it pretty much confirms what kopite7kimi leaked.
looks like n31 might be in serious trouble.

 

https://www.google.com/search?q=RX+7900+XTX+gaming+benchmarks&sourceid=chrome&ie=UTF-8

 

can you be more specific ?
all those link are either amd's or complete turds like geekbench

 

Here bro ( I do not know if they are real :/ , take it with a teaspoon of salt ) -

 

they are not

 

Which one will you get? XT or XTX?

 

What I’ve heard as well. The 7900xtx is between the 4080 and 4090. Rt performance was around 3090Ti level.

 

where ?

 

see what kopite7kimi leaked
https://forums.guru3d.com/threads/f...o-geforce-rtx-4080.445838/page-3#post-6081349

drivers affecting a synthetic benmchark that much is very improbable.
I'm seeing more and more poeple pointing to some engineering flaw related to power.

 

Fixed for you bro

" I'm seeing more and more poeple pointing to some engineering flaw related to power " - Where ???

 

among people on tech sites I visit, like purepc
and where did you see early drivers affecting 3dmark so much before ? mind you, fse scores are fine. actually amd usually punches above its weight in early 3dmark leaks.

 

True...

I also hope that the cards perform much better otherwise we are in trouble...

 

purepc ??? where else ???

Dude i have been testing gpu drivers for overclocking gains on benchmarks for hwbot for ~20 years on liquid nitrogen , phase change cooling , water chillers at minus 0 temps , i do have the experience to talk about drivers optimizations and failures for 3DMarks with early drivers , that's my cup of tea kapish ...

 

okay, but did you test leaks against release, and found a huge difference ?
this whole thing is unconfirmed no matter the site I quote, it's not like more people thinking this way has any impact. but follow that kopite7kimi post from dec 8th (a well-known leaker with good info) to what I wrote about those dx12 vs dx11 results. we know they decoupled shader clock from core clock too quite late into the cycle.

https://forum.beyond3d.com/threads/...-discussion-thread.63037/page-43#post-2278703

 

Yeah, but are the AI accelerators even being used?

 

If this is true, no one is paying a grand for those AMD cards.

 

Indeed, I was looking for two, but I'll sit patiently to see what the actual story is...

 

I don't understand the negative comments.
Is a faster option than NVIDIA gives us at the same price point.
If you want to spend 1000+ on a card go on NVIDIA, if you want to spend 1000 or less probably AMD is better.
Is that important who has the best card?